Prevalence of coinfections in women living with human immunodeficiency virus in Northeast Brazil

Abstract INTRODUCTION: Despite the success of antiretrovirals, human immunodeficiency virus (HIV) coinfections continue to cause mortality. We investigated the prevalence of coinfections in women with HIV/acquired immunodeficiency syndrome in Sergipe, Brazil. METHODS: We conducted a cross-sectional study. The coinfections investigated were syphilis, hepatitis B and C, toxoplasmosis, rubella, tuberculosis, and cytomegalovirus. RESULTS: Among the 435 women, 85 (19.5%) had coinfections. The most prevalent was HIV/syphilis, followed by tuberculosis, toxoplasmosis, hepatitis C, hepatitis B, and rubella. Additionally, 300 (96.2%) were seropositive for cytomegalovirus immunoglobulin G. CONCLUSIONS: Despite significant progress in the treatment for people with HIV, coinfections continued to affect this population.

In Brazil, although the prevalence of HIV is higher in men than that in women, women comprise an increasing proportion of the population affected by HIV, and approximately 4,000 women die from acquired immunodeficiency syndrome (AIDS)related illnesses annually 3 . In this context, coinfections among HIV-infected women are concerning, specifically for women in reproductive age or who are pregnant. Furthermore, serious complications or death of the women and/or fetus can be caused by worsening of the HIV infection 4 .
In this study, we investigated the prevalence of coinfections in women living with HIV/AIDS attended at a reference outpatient clinic for individuals living with HIV/AIDS in Sergipe State, Northeast Brazil. We conducted a cross-sectional study from August 2014 to November 2017 in women living with HIV/AIDS who attended the Reference Centre for Sexually Transmitted Diseases, HIV, and AIDS (CRIST/AIDS), the unique reference care center for individuals with HIV in Sergipe State, Northeast Brazil.
The sample size was calculated based on the number of HIV-positive women registered at the CRIST/AIDS in 2014 (800 women). The prevalence of the expected coinfections was 50%, confidence interval (CI) was 95%, and maximum error was 5%, which resulted in 260 women. However, at the end of the study, 435 HIV-positive women were included.
Data were collected through individual interviews with a structured form. For the viral load and CD4+ T-lymphocyte count, only the data obtained less than 1 year prior to the interview were included.
Patients diagnosed with HIV infection in the CRIST/AIDS underwent initial complementary examinations including tests for syphilis, viral hepatitis, toxoplasmosis, tuberculosis (TB), cytomegalovirus, and rubella. In this study, we opted to analyze these coinfections that were screened through routine examinations of the service itself. We retrieved information from all medical records from HIV diagnosis until the interview's date. Additionally, the Brazilian Information System for Notifiable Diseases (SINAN) database of TB was used.
Toxoplasma gondii and Rubella virus immunoglobulin M (IgM) and immunoglobulin G (IgG) antibodies were recorded; however, only the IgG test results were considered for cytomegalovirus because IgM test had been requested less frequently in the clinic. Syphilis was screened using treponemal and non-treponemal tests. Positive hepatitis B and C diagnosis was established with the following serological markers: hepatitis B surface antigen (HBsAg) and hepatitis C antibody. Finally, to identify the occurrence of HIV-TB coinfection, record linkage was performed between the database of this study and the SINAN database with TB cases among women in Sergipe between 2001 and 2017.
The characteristics were analyzed using descriptive statistics. The prevalence of HIV coinfections was described as a simple proportion. Pearson's chi-squared test and Fisher's exact test were used to compare the association between coinfection and the time since HIV diagnosis. The significance level was set at 5%. For the association between predictor factors and the occurrence of coinfections, the prevalence ratio (PR) with a 95% CI was used. The data were analyzed using the Statistical Package for the Social Sciences version 20.0 (International Business Machines Corporation, Armonk, NY).
This study was approved by the Research Ethics Committee of the Federal University of Sergipe (CAAE No. 92514618.8.0000.5546) and following the Helsinki Declaration. All participants provided written informed consent. Parents or guardians provided written informed consent before enrolling their children in the study.
When comparing the data of the HIV-positive women who had at least one coinfection with those who had no coinfections, it was observed that women who self-reported to be black (PR, 1.58; 95% CI, 0.75-3.33), those exposed to HIV through sexual intercourse (PR=1.58; 95% CI, 0.80-2.73), those who had their first sexual intercourse when they were younger than 15 years old (PR=1.51; 95% CI, 0.93-2.45), or those who had sex for money (PR=1.76; 95% CI, 0.88-3.50) were more likely to have a coinfection ( Table 3).
In Sergipe, one-fifth of the participants had some types of coinfection. In individuals living with HIV/AIDS, coinfection was an expected phenomenon because HIV infection deteriorates the immune system. However, this condition can be minimized with the timely and immediate use of ART. Studies have shown a decrease in opportunistic infections and a reduction in mortality in HIV-infected individuals after the advent of ART 1 . Despite the relatively insufficient studies regarding the magnitude of coinfections among the female population, the findings in a US cohort of HIV-positive women and their children were consistent with the previously reported finding 5 . In this context, despite advancements in the treatment of people living with HIV, coinfections are still common, leading to deaths 2 . The prevalence of coinfections can vary according to the etiological agent, even within the same population. In this study, most coinfections were prevalent between 1.8 and 3.9, except for syphilis, which had a rate of 9.1. A high prevalence of syphilis in women living with HIV was also reported by a study conducted in the Amazon region of Brazil 6 .
Based on our findings, some interpretations can be offered. A possible explanation for the high prevalence of syphilis in women living with HIV in this study is due to the number of syphilis cases in the Brazilian population that have increased in recent years. Additionally, both HIV and syphilis diseases share the same risk factors, suggesting a higher likelihood of coinfection occurring, emphasizing the importance of tracking this coinfection in this subpopulation.
It is well established that risky behaviors such as needle sharing, abuse of psychoactive substances (alcohol and drugs), multiple partners, prostitution, and poor adherence to condom use increase the chances of acquiring sexually transmitted infections. In this study, more than half of the women reported not using condoms during all sexual encounters. Studies show that not using condoms is mainly due to the belief that it is unnecessary to use condoms among heterosexual HIV-  seroconcordant couples and due to gender inequality as often the women are in long-term and oppressive relationships, where negotiating condom use with the partner can be difficult 7 .
Despite not being the most prevalent coinfection in this study, TB remains an important health problem. A national population-based study 8 using probabilistic linkage technique found an estimated 6.3% of TB prevalence in women living with HIV between 2011 and 2014, which was higher than the findings of this study. This suggests that the prevalence of TB/HIV coinfection varies widely among Brazilian regions.
In Sergipe, the relatively high prevalence of TB in women living with HIV may be the result of the adherence to Brazilian public policies on HIV and TB control, which include timely use of ART, intensive screening for latent TB infection, and early diagnosis and immediate treatment of TB with chemoprophylaxis, preventing the development of active TB infection. In fact, Sergipe is considered as one of the Brazilian states with the highest rates of ART use in coinfected individuals during the treatment of TB 9 .
The effect of ARV on the survival of individuals living with HIV also confirmed the development of chronic hepatitis caused by viruses B and C in this population, specifically in developing regions and in regions where there is a high endemicity of HIV infection. In African areas, studies show that the overall prevalence of HBsAg in adult women can range from 7% to 14% 10,11 . A Brazilian national study found that women infected with HIV were less likely coinfected with hepatitis B or C 12 .
Other findings worth noting were the high frequencies of seropositivity for cytomegalovirus, rubella, and toxoplasmosis IgG antibodies. This is very relevant, specifically for cytomegalovirus and toxoplasmosis because there is a risk of latent reactivation for these infections 13,14 . However, the high seropositivity for the rubella IgG antibody may correspond to the immunity acquired through vaccination or prior infection, which may explain the low prevalence of active infection in our study.
There have been few published studies regarding active rubella-HIV coinfection; however, several studies have analyzed the seroprevalence of the IgG antibody to evaluate the response to the rubella vaccine in individuals living with HIV, which showed similar and consistent results with our study for rubella IgG positivity. One of them identified that 89.3% were seropositive for rubella IgG and 10.69% were susceptible to coinfection 15 .
It is clear that it is not only important to understand the epidemiological characteristics of HIV-positive women but also be aware of the proportion of coinfections in this population group. However, this study has some limitations. Although we conducted face-to-face interviews, clinical information was completed from medical records and surveillance databases. Second, it was difficult to find the records of all the test results 6/6 for coinfections. Conversely, this missing information occurred randomly; hence, the principal consequence was loss of power. In this sense, it was assumed that these tests may not have been performed or their results were not included in the medical records, which may characterize failures in the coinfection screening protocol. Third, this study may not have been able to determine the true magnitude of the prevalence of these coinfections in the entire female population living with HIV in the Sergipe State because most of the women evaluated were routinely followed up and reported to be using ART. However, the final sample corresponded to more than half of the women enrolled in the service and was almost double the calculated sample size. Despite these limitations, this study determined the prevalence rates and important characteristics that can guide the care provided and the planning of activities for women living with HIV.
In conclusion, in Sergipe, one-fifth of the women living with HIV had some types of coinfection, with prevalence ranging from 1.8 to 3.9, except for syphilis, which had the highest rate of 9.1. Additionally, high seroprevalence for IgG antibody for cytomegalovirus, rubella, and toxoplasmosis was identified. Our results provide potential strategies to improve control programs for HIV by targeting interventions to population, with strengthening of public policies for the prevention, control, diagnosis, and treatment of coinfections in HIV-positive women with different characteristics.